NIFS/cfDNA Testing Does Not Detect 95–99% of Genetic Variants Found by Amniocentesis or CVS — That Figure Applies Only to Down Syndrome

The claim is that NIFS — non-invasive foetal sequencing via cell-free DNA (cfDNA) — identifies 95–99% of the genetic variants that invasive procedures like amniocentesis and chorionic villus sampling (CVS) can find. The verdict is partially false. The percentage range is real, but it describes performance on a narrow target, not the full landscape of foetal genetic abnormalities.

The strongest evidence against the broad claim comes from Srebniak et al., published in Ultrasound in Obstetrics & Gynecology in 2016. That cohort study found cfDNA missed roughly 50% of clinically relevant chromosomal abnormalities detectable by invasive prenatal diagnosis combined with chromosomal microarray. The reason is structural: cfDNA does not reliably detect microdeletions, microduplications, or balanced rearrangements — categories of abnormality that microarray-enhanced amniocentesis and CVS routinely catch. Wapner et al. in the New England Journal of Medicine (2012) reinforced this gap, showing that chromosomal microarray via invasive testing identified clinically significant copy-number variants in 6% of fetuses with structurally abnormal ultrasounds but a normal karyotype — variants cfDNA would not have flagged at all.

To steelman the claim: the 95–99% figures are not fabricated. Bianchi et al. (NEJM, 2012) reported 100% sensitivity for trisomy 21 and 97.2% for trisomy 18 in a high-risk cohort of 2,882 women. Norton et al. (NEJM, 2015) confirmed 100% sensitivity for both trisomies 21 and 18 across a general obstetric population of 15,841 women. The ACMG's 2016 position statement, authored by Gregg et al. in Genetics in Medicine, confirms cfDNA sensitivity exceeds 99% for trisomy 21 specifically. These are genuine, peer-reviewed findings from large, well-designed studies. The problem is not the numbers — it is the scope they are being asked to cover.

Both Norton et al. and ACOG Practice Bulletin No. 226 (2020) explicitly state that cfDNA does not detect the majority of chromosomal abnormalities, single-gene disorders, or structural variants identifiable by invasive testing. ACOG is unambiguous: a normal cfDNA result does not rule out foetal abnormalities. The test screens primarily for trisomies 21, 18, and 13, plus sex chromosome aneuploidies. It is a screening tool, not a diagnostic one. Amniocentesis and CVS with microarray cover a categorically broader spectrum — including the conditions cfDNA was never designed to find.

What is genuinely true: for the specific conditions cfDNA targets, it performs exceptionally well and can meaningfully reduce the need for invasive procedures in low-risk pregnancies. That is a real clinical advance. What is false is the leap from "detects nearly all Down syndrome cases" to "detects nearly all genetic variants." That leap erases an entire category of abnormalities and misrepresents the test's clinical scope in a way that could lead patients to forgo diagnostic procedures they actually need.

The manipulation pattern here is a missing denominator. Sensitivity statistics are always condition-specific — they measure how well a test finds a particular target, not how comprehensively it surveys everything that could be wrong. When a figure like "99% sensitivity" is stripped of its denominator (trisomy 21 cases only) and reattached to a much larger denominator (all genetic variants), the statistic becomes actively misleading. When evaluating any diagnostic test claim, always ask: sensitivity for what, exactly, in which population?